org.dllearner.scripts.improveWikipedia
Class ConceptSPARQLReEvaluator

java.lang.Object
  org.dllearner.scripts.improveWikipedia.ConceptSPARQLReEvaluator

public class ConceptSPARQLReEvaluator
extends Object

Author:

Sebastian Hellmann The EvaluatedDescriptions from a fragment are validated against the SPARQLendpoint. There are different strategies, see the methods;

Constructor Summary

ConceptSPARQLReEvaluator(SPARQLTasks sparqlTasks)
Constructor using default settings

ConceptSPARQLReEvaluator(SPARQLTasks sparqlTasks, int depthOfRDFS, int sparqlResultLimit)
constructor to manually set parameters

Method Summary

List<EvaluatedDescriptionPosNeg>	reevaluateConceptsByDataCoverage(List<EvaluatedDescriptionPosNeg> descToBeReevaluated, SortedSet<String> positiveSet) Accuracy is calculated as correct positive classified over (correct positive classified + incorrect negative classified) "How many are correctly positive classified?"
List<EvaluatedDescriptionPosNeg>	reevaluateConceptsByLowestRecall(List<EvaluatedDescriptionPosNeg> descToBeReevaluated, SortedSet<String> positiveSet) Accuracy is calculated as correct positive classified over all retrieved e.g. 50 correct out of 400 retrieved (50/400)

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

ConceptSPARQLReEvaluator

public ConceptSPARQLReEvaluator(SPARQLTasks sparqlTasks)

Constructor using default settings

Parameters:

sparqlTasks -

ConceptSPARQLReEvaluator

public ConceptSPARQLReEvaluator(SPARQLTasks sparqlTasks,
                                int depthOfRDFS,
                                int sparqlResultLimit)

constructor to manually set parameters

Parameters:

sparqlTasks -

depthOfRDFS -

sparqlResultLimit -

Method Detail

reevaluateConceptsByDataCoverage

public List<EvaluatedDescriptionPosNeg> reevaluateConceptsByDataCoverage(List<EvaluatedDescriptionPosNeg> descToBeReevaluated,
                                                                         SortedSet<String> positiveSet)

Accuracy is calculated as correct positive classified over (correct positive classified + incorrect negative classified) "How many are correctly positive classified?" e.g. 50 individuals of a 60-individual Category (50/60)

Parameters:

positiveSet -

reevaluateConceptsByLowestRecall

public List<EvaluatedDescriptionPosNeg> reevaluateConceptsByLowestRecall(List<EvaluatedDescriptionPosNeg> descToBeReevaluated,
                                                                         SortedSet<String> positiveSet)

Accuracy is calculated as correct positive classified over all retrieved e.g. 50 correct out of 400 retrieved (50/400)

Parameters:

positiveSet -